Posted: 2/2002

Trading Desk

Regular or Premium: What Are You Buying?

By Steven J. Parrish

How do you ensure that service quality is where you expect it to be or, more importantly, that it meets the minimum standards that you paid for?

Take something as mundane as the gasoline that you put into your car. No doubt you’ve heard jokes about regular and premium gasoline being pumped from a single tank. At least one gasoline supplier differentiates its premium offering by color — touting it as “crystal clear.”

Carriers face a similar problem in guaranteeing the quality of the bandwidth they purchase? Unfortunately, the color of light in the fiber will not give you a clue regarding the quality of the bits passing through the cable.

The technical organizations that defined the standards for today’s Synchronous Optical Networks (SONET) and Synchronous Digital Hierarchy (SDH) networks thought carefully about service quality when they defined how these networks would operate. The American National Standards Institute (ANSI) defines in-service digital transmission monitoring in ANSI document T1.231-1997. The International Telecommunications Union (ITU) defines similar quality and availability targets for SDH and Plesiochronous Digital Hierarchy (PDH) networks in G.826 (02/99). Together, these documents specify how quality may be defined and measured for the most broadly deployed broadband pipes in the world today.

We will focus on SONET in the remainder of this article as the underlying concepts are quite similar for SDH. SONET specifies a relatively large number of parameters and techniques for the uniform and consistent measurement of bandwidth performance. The three metrics that have gained prominent importance for carriers and customers are errored seconds, severely errored seconds and unavailable seconds.

If 10 contiguous SES are experienced, they are then transitioned to the final category, unavailable seconds (UAS). The SONET pipe is unavailable at the onset of 10 contiguous SES, and it is not considered available again until the onset of 10 contiguous seconds that are error free. Given the terminal nature of an unavailable second, it is easy to understand why this metric is used extensively in SLAs.

Now that we know what a UAS is, what would cause one to occur?

It is usually a major event with some typical cases being a cut cable (hard for light to pass through a shovel), a faulty transmitting laser or a burned out receiver. With such events, it is easy to see why an unavailable second would occur.

It also is interesting to note that in each of these events, a protected SONET circuit with route diversity would protect the underlying pipe and prevent the occurrences of any UAS for the consumer.

Since the criteria for ES, SES and UAS are defined clearly by standards documents, it is relatively easy to build associated equipment that includes the capability to measure and report on these parameters. SONET-based fiber optic terminals, optical switches and multiplexers typically include the ability to track, record and report on these performance metrics.

They do this by examining the incoming SONET signal and by reading the associated overhead administrative bits. Manufacturers may differentiate on the number of additional parameters they encompass, how long they hold historical data and to what level of granularity they go, but the basics usually are covered.

An element manager or operational support system (OSS) provides an interface between the SONET equipment and the human operators for reporting and alerting the resultant quality of service (QoS) information (See Figure).

UAS typically is specified in the SLA between a carrier and its customer. For example, the customer may specify that the circuit is required to be available 99.6 percent of the time. This would be measured across the SONET path using the SONET devices. The element manager would monitor the overall performance and report QoS information to the carrier. The carrier may set a threshold alarm if the UAS rise above 0.4 percent in this case, since that matches the customer’s SLA requirements.

Measuring QoS in SONET
Source: LighTrade Inc. (www.lightrade.com)

QoS and Trading

A recent development in the telecom industry is the concept of bandwidth trading. Bandwidth trading typically uses standardized agreements between buyers and sellers to reduce the contract negotiation period between the counterparties.

Bandwidth trading uses pooling points as depots to drop off and receive bandwidth that has been traded. So far, the trading efforts primarily have focused on SONET and SDH offerings due to the ubiquity of these technologies.

QoS plays a large role in bandwidth trading as the criteria are used to ensure the buyer receives the quality it purchased. The seller also relies on the QoS parameters to prove its adherences to the contract specifications. The agreement between the buyer and seller may include significant damage clauses so the overall quality of the delivered bandwidth is an important issue.

The trading community also focuses on UAS as the primary measure of bandwidth quality. This is convenient as the measurement, tracking and reporting process is already in place. The pooling points use sophisticated SONET equipment and systems to track and report QoS results in an unbiased manner to trading counterparties.

A lot of planning, engineering and money has gone into the construction and deployment of the world’s SONET and SDH networks. QoS was integral to the basic design and is leveraged by bandwidth buyers and sellers. It may not be as easy as looking at the color of the gas going into your tank, but the quality of SONET and SDH bandwidth is readily available using the same equipment used to provide these ubiquitous networks.

Steve J. Parrish is senior vice president networks at LighTrade Inc., which provides neutral, fully automated bandwidth pooling point services that deliver connectivity and provisioning services, including QoS monitoring and reporting, to metro and long-haul carriers, enterprise customers and intermediaries.